AIThe Registerabout 3 hours ago
AI struggles to patch vulns without adult supervision
3 min read
Autonomous AI patches from ChatGPT 5.5 and Claude Opus 4.8 cleanly fixed vulnerabilities only 26.0 percent of the time in 1Password's Off-by-1 Labs tests. Researchers propose the FLAWED acronym and released a patch evaluation harness for organizations.
Level
Hype check
Tap to vote and see what everyone thinks.
Summary by ByteBrief
